Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Earthquake Magnitude
Earthquake magnitude scales: Logarithmic measure of earthquake size
– amplitude of biggest wave: Magnitude 6 quake 10 * Magnitude 5
– energy: Magnitude 6 quake is about 32 * Magnitude 5
Richter Magnitude is calculated from the maximum amplitude of waves recorded on a seismogram, and distance to the earthquake.
10ML = 10
(logA+2.76log!!2.48)
10ML = 10
(logA)10
(2.76log!!2.48)
this is a constant, we can call it “k”.
10ML = A k
How does the maximum shaking amplitude A of a M6 quake compare to a M4 quake (same hypocenter, same seismograph location?
32ML = E k
How does the energy (moment) of a M6 quake compare to that of a M4 quake?
By the same token, you can easily compare the energy release (“seismic moment”) of different magnitude earthquakes:
A quick method (pre electronic calculators), involved using a graphical construct (called a Nomogram) that takes care of the mathematics by constructing the axes in a particular fashion.
The Nomogram allows one to compute the magnitude by plotting the distance between the quake and observatory on the left axis, and the seismometer deflection in millimetres on the right axis. (The amplitude of the deflections are what would have been recorded by a Wood- Anderson seismometer – the actual ground motions have been multiplied by 2000 which is the amplification of the Wood - Anderson seismometer at these frequencies). The points on the left and right axes are connected by a straight line, and the intersection on the middle axis is the earthquake magnitude.
*whichever’s biggest at a period of 0.8s (typically the S wave), and ALSO always using a Wood-Anderson seismograph (or converting the amplitude so the seismogram looks just like one from a WA seismograph)
N. Pinter
Mb and ML are inadequate for large earthquakes
N. Pinter
M 9.2 2004 Sumatra Earthquakemeasured in Victoria BC
local ground effects, etc.
Earthquake magnitude scales
Mw is best for large earthquakes
Mw is calculated from the earthquake energy release, which can be done with many different kinds of data, such as very long-period surface wave recordings from broadband seismometers and even GPS measurements of permanent ground displacement
C. Ammon
s is the slip (m)
A is the area of the fault that slipped in the earthquake (m )2
is that SAME constant (“shear modulus” or “rigidity”) that appeared in the seismic wave speed equations on Monday. Pascals, i.e., Newtons / m )
µ
2
Seismic moment(Newton m)
Energy released by an earthquake (seismic moment)
Mo = A µ s
To get a number that looks like a Richter magnitude, we use this equation:
= unwieldy large number, for example, “1021 Newton meters”
C. Ammon
Moment magnitude Mw comes from seismic moment Mo
Mo = A µ s
Mw = 2/3(7 + logMo) ! 10.732
3
Mw =2
3logMo ! 6.07
Mo units here are “dyne cm”. 1 dyne cm = 10-7 Newton m.
I PREFER Newton m.
http://www.tectonics.caltech.edu/slip_history/2011_taiheiyo-oki/
Moment magnitude Mw of the 2011 Tohoku earthquake from GPS
Horizontal displacements
http://www.tectonics.caltech.edu/slip_history/2011_taiheiyo-oki/
Vertical displacements
• we know how to calculate surface displacements resulting from slip on a fault.
• here, we are given surface displacements and we want to know slip on the fault (and sometimes, where the fault is and its orientation)
• this is another example of an “inverse problem”, like the earthquake location problem. (we have lots of these in geophysics)
Letʼs say you know how to calculate how much the ground deforms if there is 1 m of slip on any 1 km
patch of a fault (you do: Okada, 1985)
elastic halfspace:Okada, 1985
layered halfspace:Wang et al., 2003
3D elastic volume:FE model
You can solve for the set of fault slip patches (and how much slip on each) that produces the coseismic displacements you
see at the ground surface
this is an inverse problem
Slip inversions from surface displacement data
•! minimize L2 norm of [misfit!! between modeled and measured! surface displacements weighted ! by meas. errors]
•! smooth slip distribution (also minimize! ! curvature of slip distribution)
•! no backward slip
result
0.0
5.0
1.0
2.0
3.0
4.0
Slip (m)
0
20
10
0
100160
6020 40
80120 140
Distance (km)
bvls method, Stark and Parker, 1995
s = (" " + # H H) (" d)2$ $ $
d = ! " si ij jjj patchesi displ. dof
-1
" = Green's functionH = curvature operator
(“weighted residual sum of squares”)
http://www.tectonics.caltech.edu/slip_history/2011_taiheiyo-oki/
Mo = A µ s
Mw =2
3logMo ! 6.07
Caltech’s estimate of slip and moment for the March 2011 Tohoku Earthquake
Mw = 9.0
My estimate of slip and moment for the 1999 Izmit, Turkey Earthquake
Sum moment for all 312 4-km “patches” The seismic moment from this is 2.3 x 1020 Newton m, which is a moment magnitude of 7.4
Mo = A µ s Mw =2
3logMo ! 6.07
29˚ 00'E 29˚ 30'E 30˚ 00'E 30˚ 30'E 31˚ 00'E 31˚ 30'E40˚ 00'N
40˚ 30'N
41˚ 00'N
41˚ 30'N
0.5 m 0
km
50
Modeled and measured surface movement: 1999 Izmit, Turkey earthquake
GPS horizontal displacements with 95% confidence ellipsesmodeled horizontal displacements
surface rupture
Moment magnitude from GPS data:M = 7.5 Izmit, Turkey earthquake w
Slip for the 2009 Haiti earthquake
Mo = 4.95 ! 1019
N m
Slip and moment for the 2009 Haiti earthquake are from an inversion of surface wave data
best fitting modelseismograms
http://www.youtube.com/watch?v=A_dWf9Lr9qE
Slip versus time from an inversion of seismic records (strong motion data, frequencies
between 0.01 and 0.2 Hz)
S.-J. Lee et al., 2011.